Give stock screening live market context with a RAG agent
A commenter built a RAG agent over a live database to supply current market context for evaluating stocks and screening.
Done with a custom-built or unnamed agent
- What they did
- A commenter said plain ChatGPT was not current enough and that cross-referencing metrics got messy, so they built their own RAG agent on top of a live database. They used it to add current market context for chat-based research and stock screening, but the thread does not describe the database, retrieval flow, or any human review steps.
- How it went
- The stated outcome was better current market context for chat and screening. No performance results, accuracy numbers, or limitations beyond stale data and awkward metric cross-referencing in standard ChatGPT were given.
- Worth knowing
- The main catch is that the source gives almost no implementation detail, so you would need to design the live database and RAG workflow yourself.
Try it yourself with Claude Code
Help me build a simple retrieval agent that pulls current market context for stock screening. Connect it to [data source, such as a market data API or a database I already have], index recent news and key metrics for [list of tickers or sector], and let me ask questions like 'which of these names have positive earnings revisions this week'. Done when I can run a query from the terminal and get an answer with the underlying sources shown.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: Reddit · Undated
Related use cases
Shop for umbrella insurance through your personal agent
Personal AI agents use Coverage Cat's Agent API/MCP to size umbrella coverage and compare carrier options with price ranges, backed by a licensed broker.
Custom agent
Track personal finances with a Claude Code assistant
The author built a Claude Code financial assistant using macro data over MCP and his own financials to track his finances, assess health and support investment decisions.
Fight consumer claims and recover refunds
The author uses Claude and its Chrome extension to file complaints and dispute charges, reporting ten wins and over $12,000 recovered, while still reviewing emails and making calls.
File insurance claims for overcharges
Tarun told an agent to file all the claims after finding overcharges and has received $200 back so far.
Custom agent
Build a portfolio plan from brokerage exports
Claude Code read a directory of brokerage CSV exports and notes, researched funds, and produced buy, stop-buying and hold lists with a tax-loss harvesting partner fund for each.
Review a landlord deposit dispute for legal violations
A tenant used ChatGPT and Gemini to review a landlord's apartment deposit dispute, identify legal violations, and avoid losing $4,200.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.